Translational Regulation of TRAIL Sensitivity

TNF-related apoptosis-inducing ligand (TRAIL) is a peptide that induces apoptosis to varying degrees in tumor cells. While TRAIL sensitivity in tumors has been linked to c-myc- and MEK/Erk-induced enhancement of caspase activation, our recent study identified a third input controlling TRAIL sensitivity, namely the Akt-mTOR pathway. We showed that instead of enhancing TRAIL sensitivity, Akt expression, acting through mTOR and the mTOR targets S6 kinase and eIF-4E, selectively enhances translation of the anti-apoptotic protein FLIP(S) and confers TRAIL resistance. In this perspective article we will discuss the linkage of the Akt-mTOR pathway to other regulators of TRAIL sensitivity, its importance in controlling a broader range of apoptotic events, its utility in predicting TRAIL responsiveness, and its potential manipulation for therapeutic benefit.

mTOR controls FLIPS translation and TRAIL sensitivity in glioblastoma multiforme cells

The tumor-selective, proapoptotic, death receptor ligand tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is a mediator of antitumor drug activity and in itself is a promising agent for the treatment of human malignancies. Like many tumors, however, glioblastoma multiforme (GBM), the most fatal form of glioma, exhibits a range of TRAIL sensitivity, and only a small percentage of GBM tumors undergo TRAIL-induced apoptosis. We here show that TRAIL resistance in GBM is a consequence of overexpression of the short isoform of the caspase-8 inhibitor, c-FLICE inhibitory protein (FLIP(S)), and that FLIP(S) expression is in turn translationally enhanced by activation of the Akt-mammalian target of rapamycin (mTOR)-p70 S6 kinase 1 (S6K1) pathway. Conversely, pharmacologic or genetic inhibition of mTOR, or the mTOR target S6K1, suppresses polyribosomal accumulation of FLIP(S) mRNA, FLIP(S) protein expression, and TRAIL resistance. In archived material from 12 human GBM tumors, PTEN status was a predictor of activation of the Akt-mTOR-S6K1 pathway and of FLIP(S) levels, while in xenografted human GBM, activation status of the PTEN-Akt-mTOR pathway distinguished the tumors inherently sensitive to TRAIL from those which could be sensitized by the mTOR inhibitor rapamycin. These results define the mTOR pathway as a key limiter of tumor elimination by TRAIL-mediated mechanisms, provide a means by which the TRAIL-sensitive subset of GBM can be identified, and provide rationale for the combined use of TRAIL with mTOR inhibitors in the treatment of human cancers.

Akt activation suppresses Chk2-mediated, methylating agent-induced G2 arrest and protects from temozolomide-induced mitotic catastrophe and cellular senescence

Pharmacologic inhibition of the DNA signal transducers Chk1 and p38 blocks G2 arrest and sensitizes glioblastoma cells to chemotherapeutic methylating agent-induced cytotoxicity. Because Akt pathway activation has been suggested to also block G2 arrest induced by DNA-damaging agents and because glioma cells frequently have high levels of Akt activation, we examined the contribution of the Akt pathway to methylating agent-induced G2 arrest and toxicity. U87MG human glioma cells containing an inducible Akt expression construct were incubated with inducing agent or vehicle, after which the cells were exposed to temozolomide and assayed for activation of the components of the G2 arrest pathway and survival. Temozolomide-treated control cells activated the DNA damage signal transducers Chk1, Chk2, and p38, leading to Cdc25C and Cdc2 inactivation, prolonged G2 arrest, and loss of clonagenicity by a combination of senescence and mitotic catastrophe. Temozolomide-treated cells induced to overexpress Akt, however, exhibited significantly less drug-induced Cdc25C/Cdc2 inactivation and less G2 arrest. Akt-mediated suppression of G2 arrest was associated not with alterations in Chk1 or p38 activation but rather with suppression of Chk2 activation and reduced recruitment of Chk2 to sites of damage in chromatin. Unlike bypass of the G2 checkpoint induced by pharmacologic inhibitors of Chk1 or p38, however, Akt-induced bypass of G2 arrest suppressed, rather than enhanced, temozolomide-induced senescence and mitotic catastrophe. These results show that whereas Akt activation suppresses temozolomide-induced Chk2 activation and G2 arrest, the overriding effect is protection from temozolomide-induced cytotoxicity. The Akt pathway therefore represents a new target for the sensitization of gliomas to chemotherapeutic methylating agents such as temozolomide.

PKB/Akt mediates radiosensitization by the signaling inhibitor LY294002 in human malignant gliomas

The phosphoinositide 3-kinase (PI3-kinase) signaling pathway is frequently aberrantly activated in glioblastoma multiforme (GM) by mutation or loss of the 3' phospholipid phosphatase PTEN. PTEN abnormalities result in inappropriate signaling to downstream molecules including protein kinase B (PKB/Akt), and mammalian target of rapamycin (mTOR). PI3-kinase activation increases resistance to radiation-induced cell death; conversely, PI3-kinase inhibition enhances the sensitivity of tumors to radiation. The effects of LY294002, a biochemical inhibitor of PI3-kinase, on the response to radiation were examined in the PTEN mutant glioma cell line U251 MG. Low doses of LY294002 sensitized U251 MG to clinically relevant doses of radiation. In contrast to LY294002, rapamycin, an inhibitor of mTOR, did not result in radiosensitization. We demonstrate that among multiple known targets of LY294002, PI3-kinase is the most likely molecule responsible for LY294002-induced radiosensitization. Furthermore, using a myristoylated PKB/Akt construct, we identified PKB/Akt as the downstream molecule that mediates the synergistic cytotoxicity between LY294002 and radiation. Thus PI3-kinase dysregulation may contribute to the notable radioresistance of GM tumors and inhibition of PKB/Akt offers an excellent target to enhance radiosensitivity.

PTEN and hypoxia regulate tissue factor expression and plasma coagulation by glioblastoma

We have previously proposed that intravascular thrombosis and subsequent vasoocclusion contribute to the development of pseudopalisading necrosis, a pathologic hallmark that distinguishes glioblastoma (WHO grade 4) from lower grade astrocytomas. To better understand the potential prothrombotic mechanisms underlying the formation of these structures that drive tumor angiogenesis, we investigated tissue factor (TF), a potent procoagulant protein known to be overexpressed in astrocytomas. We hypothesized that PTEN loss and tumor hypoxia, which characterize glioblastoma but not lower grade astrocytomas, could up-regulate TF expression and cause intravascular thrombotic occlusion. We examined the effect of PTEN restoration and hypoxia on TF expression and plasma coagulation using a human glioma cell line containing an inducible wt-PTEN cDNA. Cell exposure to hypoxia (1% O(2)) markedly increased TF expression, whereas restoration of wt-PTEN caused decreased cellular TF. The latter effect was at least partially dependent on PTEN's protein phosphatase activity. Hypoxic cells accelerated plasma clotting in tilt tube assays and this effect was prevented by both inhibitory antibodies to TF and plasma lacking factor VII, implicating TF-dependent mechanisms. To further examine the genetic events leading to TF up-regulation during progression of astrocytomas, we investigated its expression in a series of human astrocytes sequentially infected with E6/E7/human telomerase, Ras, and Akt. Cells transformed with Akt showed the greatest incremental increase in hypoxia-induced TF expression and secretion. Together, our results show that PTEN loss and hypoxia up-regulate TF expression and promote plasma clotting by glioma cells, suggesting that these mechanisms may underlie intravascular thrombosis and pseudopalisading necrosis in glioblastoma.

Convection-Enhanced Delivery of Tumor Necrosis Factor-Related Apoptosis-Inducing Ligand with Systemic Administration of Temozolomide Prolongs Survival in an Intracranial Glioblastoma Xenograft Model

Although tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is a potent activator of cell death, preferentially killing neoplastic cells over normal cells, the efficacy of TRAIL for the treatment of glioma might be limited due to cellular resistance and, importantly, poor distribution after systemic administration. TRAIL and temozolomide (TMZ) were recently shown to have a synergistic antitumor effect against U87MG glioma cells in vitro. Convection-enhanced delivery (CED) can effectively distribute TRAIL protein throughout a brain tumor mass. In this study, we evaluated CED of TRAIL, alone and in conjunction with systemic TMZ administration, for antitumor efficacy. CED of TRAIL demonstrated safe and effective distribution in both normal brain and a U87MG intracranial xenograft model. Individually, both CED of TRAIL and systemic TMZ administration prolonged survival in tumor-bearing rats. However, the combination of these two treatments was significantly more effective than either treatment alone. CED of TRAIL in conjunction with systemic TMZ treatment is a promising strategy for the treatment of malignant gliomas.

Cooperative function of Chk1 and p38 pathways in activating G2 arrest following exposure to temozolomide

OBJECT

The Chk1 and p38 mitogen-activated protein kinase (MAPK) pathways play key roles in the G2 arrest caused by exposing glioma cells to temozolomide (TMZ). Although inhibition of either pathway sensitizes glioma cells to TMZ-induced cytotoxicity, the relative contributions of these pathways to TMZ-induced G2 arrest and to TMZ resistance conferred by G2 arrest have not been defined.

METHODS

The authors pharmacologically inhibited the Chk1 and/or p38 pathways in U87MG human glioma cells prior to and/or after exposure to TMZ; thereafter, effects on the TMZ-induced G2 arrest pathway and toxicity were monitored. The p38 inhibitor SB203580 or the Chk1 inhibitor UCN-01 or their combination blocked TMZ-mediated inactivation of cdc25C and cdc2, suggesting that p38 and Chk1 pathways work cooperatively and are both necessary to inactivate cdc25C and cdc2. Consistent with this idea, the inhibition of both Chk1 and p38 pathways did not lead to greater bypass of TMZ-induced G2 arrest or greater cytotoxicity than inhibition of either pathway alone. Inhibition of p38 did not alter TMZ-induced Chk1 activation/phosphorylation and vice versa, suggesting that p38 and Chk1 do not cooperatively bring about G2 arrest by reciprocal activation/phosphorylation. The two pathways, however, are not functionally identical; the Chk1 pathway was required for both the initiation and maintenance of TMZ-induced G2 arrest, whereas the p38 pathway played a role only in the initiation.

CONCLUSIONS

The Chk1 and p38 pathways cooperate to bring about TMZ-induced G2 arrest, and the inhibition of either pathway alone is sufficient to sensitize U87MG glioma cells to TMZ-induced cytotoxicity.

Integrin beta3 overexpression suppresses tumor growth in a human model of gliomagenesis: implications for the role of beta3 overexpression in glioblastoma multiforme

alphaVbeta3 integrin complexes are overexpressed in the growing, invading margins of human glioblastoma multiforme (GBM) and in the GBM vasculature, suggesting a key role for alphaVbeta3 in GBM growth and invasion. The function of alphaVbeta3 complexes in tumor formation, however, has been challenged by studies showing that loss of alphaVbeta3 expression (via loss of beta3) in the host vasculature enhances, rather than suppresses, the growth of s.c. implanted carcinomas. To directly address the role of tumor-specific alphaVbeta3 overexpression in glioma formation, we increased alphaVbeta3 expression (via overexpression of a wild-type or constitutively activated beta3) in human astrocytes genetically modified to form anaplastic astrocytoma-like tumors (Ras cells) on intracranial injection in rats. Overexpression of beta3 selectively increased levels of alphaVbeta3 integrin complexes, but had no effect on anchorage-dependent or -independent growth in vitro. After intracranial injection, however, the Ras + beta3 cells formed fewer and smaller tumors than did Ras cells. Similarly, Ras-transformed mouse astrocytes that were derived from control animals formed smaller intracranial tumors than those derived from beta3 knockout animals. Although tumors formed by human Ras and Ras + beta3 cells were similar in blood vessel density, Ras + beta3 tumors had smaller, pericyte-depleted vessels and were significantly more hypoxic, suggesting a beta3-mediated vascular defect. The growth-suppressive actions of beta3, however, could be overcome by stimulation of pathways (Akt or vascular endothelial growth factor) commonly activated in GBM. These results show that tumor-specific alphaVbeta3 overexpression has growth-suppressive effects in gliomas, but that these deleterious effects are mitigated by alterations common to alphaVbeta3-overexpressing GBM.

The p38 mitogen-activated protein kinase pathway links the DNA mismatch repair system to the G2 checkpoint and to resistance to chemotherapeutic DNA-methylating agents

Although human cells exposed to DNA-methylating agents undergo mismatch repair (MMR)-dependent G(2) arrest, the basis for the linkage between MMR and the G(2) checkpoint is unclear. We noted that mitogen-activated protein kinase p38alpha was activated in MMR-proficient human glioma cells exposed to the chemotherapeutic methylating agent temozolomide (TMZ) but not in paired cells made MMR deficient by expression of a short inhibitory RNA (siRNA) targeted to the MMR protein Mlh1. Furthermore, activation of p38alpha in MMR-proficient cells was associated with nuclear inactivation of the cell cycle regulator Cdc25C phosphatase and its downstream target Cdc2 and with activation of the G(2) checkpoint, actions which were suppressed by the p38alpha/beta inhibitors SB203580 and SB202590 or by expression of a p38alpha siRNA. Finally, pharmacologic or genetic inhibition of p38alpha increased the sensitivity of MMR-proficient cells to the cytotoxic actions of TMZ by increasing the percentage of cells that underwent mitotic catastrophe as a consequence of G(2) checkpoint bypass. These results suggest that p38alpha links DNA MMR to the G(2) checkpoint and to resistance to chemotherapeutic DNA-methylating agents. The p38 pathway may therefore represent a new target for the development of agents to sensitize tumor cells to chemotherapeutic methylating agents.

Overexpression of vascular endothelial growth factor isoforms drives oxygenation and growth but not progression to glioblastoma multiforme in a human model of gliomagenesis

Vascular endothelial growth factor (VEGF) is thought to promote tumor growth and angiogenesis. Whereas VEGF is up-regulated in only a portion of anaplastic astrocytoma (AA), it is overexpressed in most glioblastoma multiforme (GBM), and the level of expression is correlated with grade of glioma. To explore the possibility that VEGF may act as a driving force in the progression of AA to GBM, the VEGF isoforms VEGF(121) and VEGF(165) were overexpressed in genetically modified, mutant H-Ras-transformed human astrocytes that on intracranial implantation form AA-like tumors. The ability of the VEGF isoforms to stimulate growth, angiogenesis, oxygenation, and the formation of necrotic GBM-like tumors was then monitored. The parental mutant H-Ras-modified astrocytes expressed four times more endogenous VEGF than normal human astrocytes, but on intracranial implantation formed hypovascular, hypoxic, small AA-like tumors. Whereas these modest levels of VEGF overexpression were insufficient to drive oxygenation and GBM formation, an additional 8-fold increase in VEGF expression mediated by retroviral infection with constructs encoding either VEGF (121) or VEGF (165) resulted in cells which, after intracranial implantation, formed tumors that were larger, more vascular, and better oxygenated than those formed by the mutant H-ras parental cells. However, the tumors formed by the cells expressing exogenous VEGF (121) or VEGF (165) retained the phenotype of AA, lacking areas of necrosis that are the hallmark of the GBM phenotype. These results suggest that whereas the VEGF(121) and VEGF(165) isoforms can contribute to glioma vascularization, oxygenation, and growth, they do not in and of themselves drive the formation of the GBM phenotype.

Delayed repletion of O6-methylguanine-DNA methyltransferase resulting in failure to protect the human glioblastoma cell line SF767 from temozolomide-induced cytotoxicity

OBJECT

Temozolomide (TMZ)-induced O6-methylguanine (MG) DNA lesions, if not removed by MG-DNA methyltransferase (MGMT), mispair with thymine, trigger rounds of futile mismatch repair (MMR), and in glioma cells lead to prolonged G2-M arrest and ultimately cell death. Depletion of MGMT by O6-benzylguanine (BG) sensitizes tumor cells to TMZ, and this combination is currently used in clinical trials. The use of the TMZ+BG combination in gliomas, however, is complicated by the prolonged TMZ-induced G2-M arrest, which may delay activation of poorly defined cell death pathways and allow for MGMT repletion and reversal of toxicity.

METHODS

To address these issues, the actions of TMZ were monitored in DNA MMR-proficient SF767 glioma cells depleted of MGMT by BG, and in cells in which BG was removed at various times after TMZ exposure. In MGMT-depleted cells, TMZ exposure led to DNA single-strand breaks and phosphorylation of cdc2, followed by G2-M arrest, induction of p53/p21, and DNA double-strand breaks. Although DNA single-strand breaks, phosphorylation of cdc2, and G2-M arrest could be reversed by repletion of MGMT up to 5 days after TMZ exposure, TMZ-induced cytotoxicity could only be prevented if MGMT was replenished within 24 hours of the onset of G2-M arrest, and before the creation of DNA double-strand breaks.

CONCLUSIONS

These results indicate that although SF767 glioma cells undergo a prolonged G2-M arrest in response to TMZ, their ability to escape TMZ-induced cytotoxicity by MGMT repletion is limited to an approximately 24-hour period after the onset of G2-M arrest.

Defined human cellular systems in the study of glioma development

The creation and characterization of permanent cell lines derived from primary human gliomas in the 1960s gave scientists access to unlimited, renewable material in which to study the development of brain tumors. These cells, however, were already tumorigenic and selected for growth in culture, limiting the amount of information that could be gathered about the events that led to the formation of their tumors of origin. In response to these limitations, investigators moved to the study of primary tumors to identify in a correlative fashion the lesions important in tumor formation, and to the use of animal models to gain information about the transformation process. While these approaches have been unarguably successful, they too are limited by their correlative nature (in the case of cytogenetic studies) and their potential lack of direct relevance to human cancer (in the case of mouse models). Recent developments in the isolation and manipulation of human cells have allowed investigators a fresh chance to address questions about brain tumors in a direct fashion. The present review serves as a summary of important recent findings derived from the use of defined human cells in the study of gliomas.

Akt pathway activation converts anaplastic astrocytoma to glioblastoma multiforme in a human astrocyte model of glioma

Human malignant gliomas are thought to develop as the result of stepwise accumulations of multiple genetic alterations. Recently, we showed that E6/E7-mediated inactivation of p53/pRb, ras pathway activation (initiated by expression of mutant H-Ras), and expression of human telomerase reverse transcriptase (hTERT) in combination converted normal human astrocytes into cells that formed intracranial tumors resembling human anaplastic astrocytoma (AA). In this study, we created human astrocytes that, in addition to expressing E6/E7, hTERT, and Ras, also expressed a constitutive activated form of Akt intended to mimic the Akt activation noted in grade IV glioblastoma multiforme (GBM). Although these cells grew no differently than astrocytes expressing E6, E7, and H-Ras in vitro or in the first 28 days following s.c. implantation, they ultimately formed tumors four to six times larger than those formed by the E6/E7/hTERT/Ras cells. Unlike the poorly vascularized, necrosis-free AA formed by E6/E7/hTERT/Ras cells, the tumors formed by s.c. or intracranial injection of Akt-expressing cells had large areas of necrosis surrounded by neovascularization and were consistent in appearance with grade IV human GBM. These results show that activation of the Akt pathway is sufficient to allow conversion of human AA to human GBM.

Abrogation of the Chk1-mediated G(2) checkpoint pathway potentiates temozolomide-induced toxicity in a p53-independent manner in human glioblastoma cells

Temozolomide (TMZ) produces O(6)-methylguanine in DNA, which in turn mispairs with thymine, triggering futile DNA mismatch repair (MMR) and ultimately cell death. We found previously that in p53-proficient human glioma cells, TMZ-induced futile DNA MMR resulted not in apoptosis but rather in prolonged, p53- and p21-associated G(2)-M arrest and senescence. Additionally, p53-deficient cells were relatively more TMZ resistant than p53-deficient glioma cells, which underwent only transient G(2)-M arrest before death by mitotic catastrophe. These results suggested that prolonged G(2)-M arrest might protect cells from TMZ-induced cytotoxicity. In the present study, we therefore focused on the mechanism by which TMZ induces G(2)-M arrest and on whether inhibition of such G(2)-M arrest might sensitize glioma cells to TMZ-induced toxicity. U87MG glioma cells treated with TMZ underwent G(2)-M arrest associated with Chk1 activation and phosphorylation of both cdc25C and cdc2. These TMZ-induced effects were inhibited by the Chk1 kinase inhibitor UCN-01. Although not in itself toxic, UCN-01 increased the cytotoxicity of TMZ 5-fold, primarily by inhibiting cellular senescence and increasing the percentage of cells bypassing G(2)-M arrest and undergoing mitotic catastrophe. In addition to enhancing TMZ-induced cytotoxicity in p53-proficient cells, UCN-01 also blocked TMZ-induced Chk1 activation and transient G(2)-M arrest in p53-deficient U87MG-E6 cells and similarly enhanced TMZ-induced mitotic catastrophe and cell death. Taken together, these results indicate that Chk1 links TMZ-induced MMR to G(2)-M arrest. Furthermore, inhibition of the cytoprotective G(2) arrest pathway sensitizes cells to TMZ-induced cytotoxicity and may represent a novel, mechanism-based means of increasing TMZ efficacy in both p53 wild-type and p53 mutant glioma cells.

Formation of intracranial tumors by genetically modified human astrocytes defines four pathways critical in the development of human anaplastic astrocytoma

The formation of human malignant gliomas is thought to involve the accumulation of multiple genetic alterations. To define the function of specific alterations in glioma formation, we serially introduced genetic alterations functionally equivalent to those noted in human malignant gliomas into normal human astrocytes (NHAs). We then monitored the ability of each of these alterations to contribute to the growth of otherwise genetically stable NHAs into intracranial malignant gliomas. Using this model, we show that expression of human telomerase catalytic component (hTERT), but not E7-mediated inactivation of pRb or E6/E7-mediated inactivation of p53/pRb, was sufficient to initiate the tumorigenic process by circumventing cellular senescence in astrocytes. hTERT expression, even in combination with inactivation of p53/pRb, did not transform astrocytes. These alterations together, however, cooperated with ras pathway activation (initiated by expression of mutant H-Ras), but not with phosphatidylinositol 3-kinase pathway activation (initiated by expression of myristoylated Akt) or epidermal growth factor receptor activation, to allow for the formation of intracranial tumors strongly resembling p53/pRb pathway-deficient, telomerase-positive, ras-activated human grade III anaplastic astrocytomas. These results identify four pathways as key in the development of human anaplastic astrocytomas.

Dual growth arrest pathways in astrocytes and astrocytic tumors in response to Raf-1 activation

Normal human fibroblasts have been shown to undergo a p16(Ink4a)-associated senescence-like growth arrest in response to sustained activation of the Ras/Raf/MEK/ERK pathway. We noted a similar p16(Ink4a)-associated, senescence-like arrest in normal human astrocytes in response to expression of a conditional form of Raf-1. While HPV16 E7-mediated functional inactivation of the p16(Ink4a)/pRb pathway in astrocytes blocked the p16(Ink4a)-associated growth arrest in response to activation of Raf-1, it also revealed a second p21(Cip1)-associated, senescence-associated, beta-galactosidase-independent growth arrest pathway. Importantly, the p21(Cip1)-associated pathway was present not only in normal astrocytes but also in p53-, p14(ARF)-, and p16(Ink4a)/pRb-deficient high grade glioma cells that lacked the p16(Ink4a)-dependent arrest mechanism. These results suggest that normal human cells have redundant arrest pathways, which can be activated by Raf-1, and that even tumors that have dismantled p16(Ink4a)-dependent growth arrest pathways are potentially regulated by a second p21(Cip1)-dependent growth arrest pathway.

p53 effects both the duration of G2/M arrest and the fate of temozolomide-treated human glioblastoma cells

Temozolomide (TMZ) is a DNA-methylating agent that has recently been introduced into Phase II and III trials for the treatment of gliomas. TMZ produces O6-methylguanine in DNA, which mispairs with thymine during the next cycle of DNA replication. Subsequent futile cycles of DNA mismatch repair can lead to a p53-associated apoptotic cell death, although this mechanism has been described mostly in hematopoietic neoplasms. We studied the action of TMZ in gliomas and the role p53 might play by using U87 glioma cells that were either p53-wild-type or p53-deficient (by virtue of expression of the viral oncoprotein E6). LN-Z308 cells, in which p53 gene is deleted, were also used. p53-proficient U87 MG cells underwent a prolonged, p53- and p21(Waf1/Cip1)-associated G2-M arrest beginning 2 days after TMZ treatment. Although very few of these cells underwent apoptosis, most underwent senescence over a 10-day period. p53-deficient (E6-transfected U87 and LN-Z308) cells similarly underwent G2-M arrest in response to TMZ, but this arrest was accompanied by only minor changes in p53 or p21(Waf1/Cip1) and was reversed within 7 days of TMZ treatment in association with the appearance of cells with either 8n or subG1 DNA content. These results suggest that glioma cells respond to TMZ by undergoing G2-M arrest. p53 is not necessary for this G2-M arrest to occur but is important in the duration of G2-M arrest and in the ultimate fate of TMZ-treated cells. Therefore, the integrity of the G2-M cell cycle checkpoint may be important in the cytotoxicity of TMZ in glioma cells.

Phosphatidylserine-dependent phagocytosis of apoptotic glioma cells by normal human microglia, astrocytes, and glioma cells

Apoptotic cells display signals that trigger phagocytic removal by macrophages or neighboring cells. To better understand the signals triggering phagocytosis of apoptotic glioma cells, and to identify the cells that might be involved in the phagocytic process, U-251 MG glioma cells were made apoptotic by etoposide (25 microg/ml) treatment and were incubated with normal human astrocytes (NHA), glioma cells, or microglia. Extent of phagocytosis was assessed by an in vitro phagocytosis assay. After 3 h of incubation with apoptotic cells, phagocytes tested were washed to remove nonengulfed cells, then fixed, stained, and counted to determine phagocytosis index (PI). NHA, glioma cells, and microglia all phagocytosed apoptotic, but not nonapoptotic, glioma cells. Microglia, however, had a PI approximately 4-fold higher than did either NHA or glioma cells. Binding of phosphatidylserine (PS) on apoptotic glioma cell membranes by annexin-V inhibited phagocytosis by 90% in both microglia and NHA. The activity of an enzyme (scramblase) that moves PS from the inner cell membrane to the outer cell membrane was also increased in apoptotic glioma cells. These results suggest that a variety of cells present in and near gliomas in vivo can remove glioma cells in a PS-dependent scramblase-mediated fashion. Manipulation of scramblase and/or PS exposure in glioma cells may therefore be a means of triggering phagocytic removal of glioma cells.

Presetting of chromatin structure and transcription factor binding poise the human GADD45 gene for rapid transcriptional up-regulation

GADD45 has been suggested to coordinate cell cycle regulation with the repair of DNA damage following ionizing radiation (IR). Although the GADD45 gene is transcriptionally up-regulated in response to IR, alterations in in vivo transcription factor (TF) binding or chromatin structure associated with up-regulation have not been defined. To understand how chromatin structure might influence TF binding and GADD45 up-regulation, key regulatory regions of the gene were identified by in vivo DNase I hypersensitivity (HS) analysis. Chromatin structure and in vivo TF binding in these regions were subsequently monitored in both non-irradiated and irradiated human ML-1 cells. In non-irradiated cells expressing basal levels of GADD45, the gene exhibited a highly organized chromatin structure with distinctly positioned nucleosomes. Also identified in non-irradiated cells were DNA-protein interactions at octamer binding motifs and a CCAAT box in the promoter and at consensus binding sites for AP-1 and p53 within intron 3. Upon irradiation and a subsequent 15-fold increase in GADD45 mRNA levels, neither the chromatin structure nor the pattern of TF binding in key regulatory regions was altered. These results suggest that the GADD45 gene is poised for up-regulation and can be rapidly induced independent of gross changes in chromatin structure or TF binding.

Confluence-induced alterations in CpG island methylation in cultured normal human fibroblasts

Growth constraint of bacterial and human cells has been shown to trigger genetic mutation. We questioned whether growth constraint might also trigger epigenetic mutation in the form of CpG island methylation. Logarithmically growing normal human fibro-blasts (NHF) displayed little (0-15%) CpG methylation in select regions of three CpG islands [estrogen receptor (ER), E-cadherin (ECAD) and O (6)-methylguanine-DNA methyltransferase (MGMT)] examined. NHF grown to and left at confluence for 2-21 days showed little (<10%) CpG methylation in the ER and ECAD CpG islands. These confluent, growth-arrested cells, however, displayed extensive ( approximately 50%) methylation of the MGMT CpG island. CpG methylation in the MGMT CpG island was not associated with cellular senescence. The methylation was, however, heritable, but not permanent, as the level of CpG methylation in the MGMT CpG island of cells 4 population doublings following replating after confluence were no different from those in confluent cultures, but returned to levels noted in logarithmically growing cells by 10 population doublings following replating. These results suggest that growth constraint can trigger transient epigenetic change even in normal non-senescent human cells.

Understanding and manipulating O6-methylguanine-DNA methyltransferase expression

O6-Methylguanine-DNA methyltransferase (MGMT) is a DNA repair protein that transfers methyl and alkyl lesions from the O6 position of guanine to a cysteine in its structure. The ability of MGMT to also remove precytotoxic O6-alkylguanine lesions induced by chemotherapeutic chloroethylnitrosoureas has made down-regulation of MGMT expression the key component in strategies designed to sensitize tumors to the cytotoxic potential of chloroethylnitrosoureas. The study of how to regulate MGMT expression at the gene, mRNA, and protein levels has contributed not only to the development of effective inhibitors of MGMT action, but also, in a broader sense, to a better understanding of gene regulation and protein structure/function.

Aberrant silencing of the CpG island-containing human O6-methylguanine DNA methyltransferase gene is associated with the loss of nucleosome-like positioning

Tumor-associated aberrant silencing of CpG island-containing genes has been correlated with increased cytosine methylation, a "closed" chromatin structure, and exclusion of transcription factor binding in the CpG island/promoter regions of affected genes. Given the lack of understanding of what constitutes a closed chromatin structure in CpG islands, however, it has been difficult to assess the relationship among cytosine methylation, chromatin structure, and inappropriate gene silencing. In this study, nuclease accessibility analysis was used to more clearly define the chromatin structure in the CpG island of the human O6-methylguanine DNA methyltransferase (MGMT) gene. Chromatin structure was then related to in vivo DNA-protein interactions and cytosine methylation status of the MGMT CpG island in human glioma cells varying in MGMT expression. The results of these studies indicated that the "open" chromatin structure associated with the MGMT CpG island in MGMT+ cells consisted of an approximately 250-bp transcription factor-binding, nuclease-accessible, nucleosome-free region of DNA, whose formation was associated with at least four flanking, precisely positioned nucleosome-like structures. In MGMT- cells, this precise nucleosomal array was lost and was replaced by randomly positioned nucleosomes (i.e., the closed chromatin structure), regardless of whether methylation of the CpG island was spread over the entire island or limited to regions outside the transcription factor binding region. These results suggest that CpG islands facilitate the expression of housekeeping genes by facilitating nucleosomal positioning and that the conditions that alter the formation of this array (such as perhaps methylation) may indirectly affect CpG island-containing gene expression.

Methylation of discrete regions of the O6-methylguanine DNA methyltransferase (MGMT) CpG island is associated with heterochromatinization of the MGMT transcription start site and silencing of the gene

O6-Methylguanine DNA methyltransferase (MGMT) repairs the mutagenic and cytotoxic O6-alkylguanine lesions produced by environmental carcinogens and the chemotherapeutic nitrosoureas. As such, MGMT-mediated repair of O6-alkylguanine lesions constitutes a major form of resistance to nitrosourea chemotherapy and makes control of MGMT expression of clinical interest. The variability of expression in cell lines and tissues, along with the ease with which the MGMT phenotype reverts under various conditions, suggests that MGMT is under epigenetic control. One such epigenetic mechanism, 5-methylation of cytosines, has been linked to MGMT expression. We have used an isogenic human multiple myeloma tumor cell line model composed of an MGMT-positive parent cell line, RPMI 8226/S, and its MGMT-negative variant, termed 8226/V, to study the control of MGMT expression. The loss of MGMT activity in 8226/V was found to be due to the loss of detectable MGMT gene expression. Bisulfite sequencing of the MGMT CpG island promoter revealed large increases in the levels of CpG methylation within discrete regions of the 8226/V MGMT CpG island compared to those in 8226/S. These changes in CpG methylation are associated with local heterochromatinization of the 8226/V MGMT transcription start site and provide a likely mechanism for the loss of MGMT transcription in 8226/V.

Methylation of CpG island transcription factor binding sites is unnecessary for aberrant silencing of the human MGMT gene

Aberrant transcriptional inactivation of the non-X-linked human O-6-methylguanine DNA methyltransferase (MGMT) gene has been associated with loss of open chromatin structure and increases in cytosine methylation in the Sp1-binding region of the 5'-CpG island of the gene. To examine the necessity of these events for gene silencing, we have isolated and characterized a subline of human MGMT+ T98G glioma cells. The subline, T98Gs, does not express MGMT activity or MGMT mRNA, and exhibits no in vivo DNA-protein interactions at Sp1-like binding sites in the MGMT 5'-CpG island. While the MGMT CpG island is less accessible to exogenously added restriction enzymes in T98Gs nuclei than in T98G nuclei, it is similarly methylated in both T98G and T98Gs cell lines 5' and 3' to the transcription factor binding sites, and similarly unmethylated in the region encompassing the binding sites. Inappropriate transcriptional inactivation of MGMT, therefore, does not require methylation of transcription factor binding sites within the 5'-CpG island. Rather, MGMT gene silencing and transcription factor exclusion from T98Gs MGMT CpG island binding sites is most closely associated with condensed chromatin structure, which is in turn indirectly influenced by distant sites of methylation.